In order to have more performance these days, the problem is how to put data closer to the processor (locality). This is x1000 speedup issues. Better workload balance is important but it is only x4-8 speed up issues.

Here is an example where Intel processors expose APIs for data locality in multithreading programs: https://software.intel.com/content/www/us/en/develop/documentation/cpp-compiler-developer-guide-and-reference/top/optimization-and-programming-guide/openmp-support/openmp-library-support/thread-affinity-interface-linux-and-windows.html